Six events, and what radar showed
- View
Six well-documented events from six parts of the country, each illustrating one limitation this module has spent nine lessons describing.
In this lesson you will
- Relate reflectivity, velocity and dual-pol signatures to documented events.
- Explain how range and beam height affected what radar could show in specific cases.
- Distinguish measured from estimated wind in the public record.
- Locate archived radar data for self-study.
Why these six
Each one is documented on a public NWS event page with archived data, and each one breaks a different assumption. They are spread across the country on purpose: nothing in this module is regional, and neither are its failure modes.
Bridge Creek and Moore, Oklahoma, 3 May 1999
The event that produced the highest tornado wind speed ever reliably measured: a Doppler on Wheels recorded about 301 mph, plus or minus 20, at roughly 105 feet above ground. Note the height: it took a mobile radar parked close by to measure that near the surface, and it is the sort of measurement the fixed network cannot make at all.
You will see 318 and 321 mph quoted for this tornado. Those are secondary re-analyses. The figure with error bars attached is 301 plus or minus 20 mph, and a course that teaches you to prefer numbers with stated uncertainty over rounder numbers without has taught you something more useful than the number.
Tuscaloosa and Birmingham, Alabama, 27 April 2011
A violent, highly visible, well-observed EF4 during the largest outbreak in decades. Its value here is a technical accident: the Birmingham radar was still single-polarisation that day. The network-wide dual-pol upgrade had begun weeks earlier and had not reached it.
Everything available that day was reflectivity and velocity: hook, debris ball, velocity couplet. The debris ball had to be inferred from reflectivity alone, exactly as lesson 5 warned, with no correlation coefficient field to confirm that what was up there was not made of water. Polarimetric analysis of that outbreak came from a university research radar rather than from the operational network. It is a useful reminder of how recent the capability is.
Joplin, Missouri, 22 May 2011
The deadliest single US tornado since 1950, with 158 fatalities, and rain-wrapped for much of its life. The service assessment finding that lesson 3 quoted comes from here: "limited scans at lowest elevation slices during this time impacted the WFO's ability to ascertain the magnitude of the tornado."
Read that alongside lesson 3. It is the reason supplementary low-level scanning exists in the form it does, and it is a case where the limiting factor was not what the radar could see but how often it was looking at the part that mattered.
El Reno, Oklahoma, 31 May 2013
Module 8 lesson 1 covers this event for its safety record. Here it is the rating that matters. A mobile polarimetric radar measured winds greater than 295 mph in sub-vortices near the surface. The official rating is EF3.
The Enhanced Fujita scale rates damage, and official ratings do not use radar-measured winds. The surveyors could not find damage supporting a higher rating, largely because the most extreme winds passed over open country. A radar measurement and a damage rating are answers to two different questions, and the rating for this event was revised during 2013 before settling where it did. Module 5 lesson 6 covers the scale properly.
The Iowa derecho, 10 August 2020
A textbook bow echo with a rear-inflow jet, and the costliest thunderstorm event on record at around 11.2 billion dollars. Lesson 5's bow echo section describes what it looked like; module 2 lesson 8 has the physics.
| Figure | What kind of number it is |
|---|---|
| 126 mph at Atkins, Iowa | Measured by an instrument |
| 140 mph near Cedar Rapids | Estimated from damage by a survey |
Both appear in the public record for the same event and they are not in conflict. Module 7 lesson 3 asks you to tag your own numbers as measured or estimated for exactly this reason: the distinction survives into the archive and people reading it years later need it.
The Quad-State tornado, 10 to 11 December 2021
A long-track nocturnal tornado family that crossed several states, with the Mayfield, Kentucky damage among the worst of it. December, at night, over a long distance.
A debris signature was present at long range from the nearest radar. Everything lesson 2 said applies at once: the beam was thousands of feet up, the sample volume was kilometers across, and the signature was detected anyway because the debris field was enormous. It was also the middle of the night in December, when module 8 lesson 5 says visual spotting is not available. This is the case where radar was doing the most work and was still, at that range and height, not an observation of the ground.
Working the archive yourself
The full-resolution radar data for every event above is public and free. Archived Level II lives in the unidata-nexrad-level2 bucket on AWS, in us-east-1, readable with --no-sign-request and no account. Real-time chunks and Level III have their own buckets alongside it.
The older
noaa-nexrad-level2 bucket was deprecated on 1 September 2025. A tutorial that points you at it is out of date, and that is a useful check on the age of any radar tutorial you find.Open-source tools will read it: Py-ART, MetPy and xradar are all freely licensed and all render the fields this module has described. Local storm reports for the same events are archived by the Iowa Environmental Mesonet, so you can put the reports beside the radar and see for yourself how well they line up. That exercise teaches this module better than this module does.
Check yourself
For the 10 August 2020 Iowa derecho, the record contains both a 126 mph figure and a 140 mph figure. What is the relationship between them?
- 126 mph was measured by an instrument; 140 mph was estimated from damage by a survey
- They are two measurements from different instruments, and the higher one superseded the lower
- The 140 mph figure is an error that was later corrected to 126 mph
- 126 mph was a sustained wind and 140 mph was the gust
Hint
Module 7 asks you to tag every number in your own reports with exactly this distinction.
Answer
126 mph was measured by an instrument; 140 mph was estimated from damage by a survey
One was measured and one was estimated from damage, and both are legitimately in the record for the same event. This is why module 7 insists you label your own numbers: the distinction has to survive into the archive for anybody reading it later to interpret it.
Sources
- NWS Warning Decision Training Division, radar and warning guidance
- NOAA JetStream, the online school for weather: radar
- Wurman et al. 2007, Bull. Amer. Meteor. Soc. 88(1), 31–46: low-level winds in tornadoes and potential catastrophic tornado impacts in urban areas
- Knupp et al. 2014, Bull. Amer. Meteor. Soc. 95(7), 1041–1062: meteorological overview of the devastating 27 April 2011 tornado outbreak
- NWS Central Region, service assessment: Joplin, Missouri tornado, 22 May 2011
- NSF Unidata, NEXRAD archive on AWS Open Data
- Helmus & Collis 2016, J. Open Res. Software 4, e25: the Python ARM Radar Toolkit